The oxidation of hydrocarbons is an important reaction in industrial organic chemistry. Thus, for example, the oxidation of cyclohexane is used commercially to produce cyclohexanol and cyclohexanone, which are important precursors in the production of nylon, whereas oxidation of alkylaromatic hydrocarbons is used in the production of phenol, a precursor in the manufacture of polycarbonates and epoxy resins.
Oxidation of hydrocarbons can be conducted using well-known oxidizing agents, such as KMnO4, CrO3, and HNO3. However, these oxidizing agents have the disadvantage of being relatively expensive, and moreover their use is accompanied by the production of unwanted coupling products which can represent disposal problems and ecological pollution.
Preferably, therefore, oxidizing agents based on peroxides or N2O are used. The cheapest oxidizing agent, however, is molecular oxygen, either in pure form or as atmospheric oxygen. However, oxygen itself is usually unsuitable for oxidizing hydrocarbons, since the reactivity of the O2 molecule, which occurs in the energetically favorable triplet form, is not sufficient.
By using redox metal catalysts it is possible to utilize molecular oxygen for oxidizing organic compounds and hence a great number of industrial processes are based on the metal-catalyzed autooxidation of hydrocarbons. Thus, for example, the oxidation cyclohexane with O2 to cyclohexanol and/or cyclohexanone proceeds with the use of cobalt salts. These industrial processes are based on a free-radical chain mechanism, in which the bi-radical oxygen reacts with a hydrocarbon free radical, with formation of a peroxy radical and subsequent chain propagation by abstraction of an H atom at a further hydrocarbon. In addition to metal salts, however, organic molecules can also act as free-radical initiators.
However, it is a disadvantage of these processes that the selectivity decreases very greatly with increasing conversion rate and therefore the processes must be operated at a very low level of conversion rate. Thus, for example, the oxidation of cyclohexane to cyclohexanol/cyclohexanone is carried out at a conversion rate of 4 to 7% so that the selectivity is 80 to 85% (Ulf Schuchardt, et. al., Applied Catalysis A: General, 211 (2001) 8).
An alternative to metal salt catalysts is the use of organic mediators, for example N-hydroxyphthalimide (NHPI). Thus, U.S. Pat. Nos. 6,852,893 and 6,720,462 describe methods for oxidizing hydrocarbon substrates by contacting the substrate with an oxygen-containing gas, in which the oxygen content is from 5 to 100% by volume, in the presence of a free radical initiator and a catalyst, typically an N-hydroxycarbodiimide catalyst, such as N-hydroxyphthalimide (NHPI). The process is conducted at a temperature between 0° C. and 500° C. and a pressure between atmospheric and 100 bar (100 and 10,000 kPaa). The molar ratio of the catalyst to the hydrocarbon substrate can range from 10−6 mol % to 1 mol %, whereas the molar ratio of free-radical initiator to the catalyst can be 4:1 or less, such as 1:1 to 0.5:1. Suitable substrates that may be oxidized by this process include cumene, cyclohexylbenzene, cyclododecylbenzene, and sec-butylbenzene.
U.S. Pat. No. 7,038,089 discloses a process for preparing a hydroperoxide from a hydrocarbon selected from a group consisting of primary hydrocarbons, secondary hydrocarbons and mixtures thereof corresponding to said hydroperoxide which comprises conducting oxidation of said hydrocarbon at a temperature in the range between 130° C. and 160° C. with an oxygen-containing gas in a reaction mixture containing said hydrocarbon and a catalyst comprising a cyclic imide compound and an alkali metal compound. Suitable hydrocarbons are said to include C4 to C20 tertiary alkanes (e.g., iso-butane, iso-pentane, iso-hexane, and the like), C7 to C20 (alkyl) aromatic hydrocarbons with 1 to 6 aromatic rings or C9 to C20 (cycloalkyl) aromatic hydrocarbons with 1 to 6 aromatic rings (e.g., xylene, cumene, cymene, ethylbenzene, diisopropylbenzene, cyclohexylbenzene, tetrahydronaphthalene (tetraline), indane, etc.), and the like. The amount of the cyclic imide compound used may be from 0.0001 to 1%, preferably from 0.0005 to 0.5%, by weight based on the reaction mixture, whereas the amount of the alkali metal compound may be from 0.000005 to 0.01%, preferably from 0.00001 to 0.005%, by weight based on the reaction mixture.
However, although current work has continued to demonstrate the utility of cyclic imides as hydrocarbon oxidation catalysts, it has also shown that their application in a commercial process requires further investigation. In particular, cyclic imides, such as N-hydroxyphthalimide, are expensive and are readily hydrolyzed under the conditions of the oxidation reaction. Since water is inherently generated in the oxidation process, loss of catalyst through hydrolysis represents a significant problem to the commercial viability of the process.
According to the invention, it has now been found that, despite the inherent generation of water in the process, the problem of catalyst loss through hydrolysis during the oxidation of hydrocarbons using cyclic imide catalysts can be significantly reduced by at least partially drying the oxygen-containing gas used in the oxidation to a water level of 0.6% by weight or less. This can be readily achieved by the simple expedient of compressing and/or cooling the oxygen-containing gas to a temperature of about 3° C. to about 20° C. and then passing the gas through a gas/liquid separator. If desired, lower levels of water can be achieved by passing the oxygen-containing gas through a molecular sieve.